Buying a camera-based safety system is unusual for most H&S teams. It touches IT, data protection, the workforce and the budget at once, and the demos are built to show the best clip. These are the questions that separate the systems in practice, written so they apply to any supplier.
Many UK sites already have CCTV over the yard, the docks and the main aisles, installed for security. Some systems use those feeds; others need their own cameras. Neither is wrong, but it changes the cost, the timescale and the disruption.
The better question is what happens where an existing camera is not good enough. Position, angle, resolution and lighting all matter. A supplier who assesses each camera before relying on it, and says plainly which ones will not work, is more useful than one who says yes to all of them.
Footage of identifiable people is personal data under UK GDPR. The ICO's guidance on video surveillance expects you to know where it goes, who processes it, and how long it is kept. If processing happens outside the UK, you need to understand the transfer arrangements.
Ask for the location of processing and storage, the retention period, and whether the supplier will support your data protection impact assessment. A camera system that analyses behaviour is very likely to need one (UK GDPR Article 35).
Connecting to a site's camera network usually means a security questionnaire, firewall changes and an IT ticket queue. That is reasonable, and it can add weeks before anyone learns whether the system is any use.
Ask whether there is a way to start without touching your network, so the first question you answer is "does this find anything worth acting on?" rather than "can we get it connected?". Ask also for the supplier's security credentials, such as Cyber Essentials.
Every system will list what it detects. Fewer will tell you what it misses, how often it flags something that turns out to be nothing, and who checks before an event reaches you.
Useful questions: which events are detected on our layout, not the demo's? What happens in the dark, in rain, at a busy dock door? Is there a person reviewing events, or does everything land in your inbox? A system that floods the safety team with false alarms soon stops being looked at, whatever its accuracy on paper.
A system that floods the safety team with false alarms soon stops being looked at.
This question decides whether the system is accepted on the floor. Employers have a legal duty to consult employees on the health and safety consequences of introducing new technology (Health and Safety (Consultation with Employees) Regulations 1996, regulation 3; Safety Representatives and Safety Committees Regulations 1977 where there is a recognised union).
Ask whether privacy masking or face blurring is available, and whether it is automatic or on request; whether access is limited and logged; whether clips are attached to events rather than browsed, and whether the system is kept separate from security and productivity monitoring. Sites that consult early, and say clearly what the footage will and will not be used for, find the conversation much easier.
A demo shows what the system can do somewhere else. A trial shows what it sees on your site. Ask whether you can start with one area, for a fixed period, with a written outcome: what was seen, by type and time of day, and what the supplier would change first. If the trial ends with nothing written down, it is hard to compare suppliers or to make the case internally.
Detection is the start. The value comes from what follows: who the event goes to, whether it becomes a record, whether an action follows, and whether anyone checks later that the fix held. HSG65's Plan, Do, Check, Act puts the weight on the Check. Ask the supplier to show you the journey of one event from clip to closed action, and how you would know the control is still working two months on.
References from your own sector and country are worth more than logos. A port, a distribution centre and a food factory share hazards but not layouts, shift patterns or cultures. Ask to speak to a safety lead at a comparable UK site, and ask them what changed after the first month, not just what the system found.
Camera-based monitoring is not the right first answer for every site, and it is worth saying so.
Monitoring earns its place where the question is what actually happens across shifts: whether the controls you have are being used, and where the near misses cluster.
| Question | Our answer |
|---|---|
| Existing cameras | RoboK SiteLens runs on the CCTV a site already has. We assess each camera first, and supply one where a zone has no usable camera, including portable units. |
| UK data | Footage is processed and stored in the UK. It is kept for 30 days as standard, and clips flagged as hazards for the length of your subscription, so you keep the evidence. Both can be changed in your contract. |
| IT | Safety Mapping starts with our own camera on its own 4G connection: mains power, no access to your network. Connecting your cameras is the next step, agreed on the review call. We provide an IT overview and data-handling documents, and RoboK is Cyber Essentials certified. |
| Detection | People in vehicle routes, near misses between vehicles and people, exclusion-zone entries, PPE where required, vehicle speed and direction. Each event has a time, a type and a clip. |
| Who sees it | The people the site nominates, on named user accounts; system access is logged. Faces can be blurred, and areas masked, on request. The safety view is separate from security and productivity systems. |
| Trial | A free two-week Safety Mapping trial on one area, ending in a written report and a review call. |
| After detection | Events build by zone and shift, so the record shows whether a change held, not just that it was made. |
| References | Ports including Bristol, Dover and Peel Ports, and logistics and manufacturing sites across the UK. |
Whether it works on your existing cameras, where footage is processed and stored, what it needs from IT, what it detects and misses, who sees the footage, how the workforce is consulted, whether you can trial it on one area with a written outcome, and what happens to an event after detection.
Yes, when it is done under UK GDPR: a lawful basis, a data protection impact assessment where the processing is likely to be high risk, transparency with the workforce, proportionate retention and security. The ICO's guidance on video surveillance sets out the expectations. Employers must also consult employees on new technology that affects their health and safety.
Not always. Some systems run on existing CCTV, subject to camera position and image quality; others need their own hardware. Ask the supplier to assess your cameras before you commit.
It depends on whether the system needs network access and new cameras. Starting on a single standalone camera is usually quicker; connecting to a site's camera network usually takes longer because of IT and security review.
Run the same test with each: one area, a fixed period, and a written report of what was seen. Then compare the reports, the false alarms, and how easily the safety team could act on what it found.
The fairest test of any of these systems is one area, two weeks, and a written report you can hold it to. That is how Safety Mapping starts. Safety Mapping: a free two-week trial on one camera.
By Hao Zheng, CEO at RoboK · Published 28 October 2026 · Last updated 28 October 2026